Multi-working-condition and multi-stage aviation equipment degradation prediction method and device

Abstract

The invention relates to a multi-working-condition and multi-stage aviation equipment degradation prediction method and device, and the method comprises the steps: fusing the performance data of aviation equipment under a multi-dimensional time-varying working condition into comprehensive health index data; performing heuristic change-point dynamic detection on the comprehensive health index data, and intercepting degradation feature data from the comprehensive health index data according to a detection result; and on the basis of a complex nonlinear Wiener degradation model, constructing a degradation track model of the aviation equipment in combination with the degradation characteristic data, and predicting the degradation trend of the equipment by using the degradation track model. According to the method, the degradation condition of the aviation equipment is accurately predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of aircraft fault prediction and health management, and particularly to a degradation prediction method and device for aviation equipment in multiple working conditions and multiple stages. Background Art

[0002] In view of the need to improve the operation safety of aviation equipment, it is necessary to predict the future health state / degradation state of aviation equipment to make decisions for the aircraft to safely perform tasks within a certain period in the future. In the actual operation of aviation equipment, due to the influence of operating conditions, the performance parameters of the equipment fluctuate greatly; in the initial stage of operation, the equipment is generally in a stable operation stage, and in the middle and later stages of operation, the equipment will enter a slow or accelerated degradation stage. Therefore, accurately evaluating the current degradation state of aviation equipment is of great significance for preventing further performance degradation of aviation equipment and can also improve the safety performance of the aircraft. The prediction of the degradation trend can accurately and timely provide relevant status information for users, which can not only effectively prevent the occurrence of equipment failures but also improve the maintenance efficiency. Summary of the Invention

[0003] The present invention provides a degradation prediction method and device for aviation equipment in multiple working conditions and multiple stages, which solves the problem of predicting the health state of aviation equipment under complex degradation processes and provides a decision-making basis for the aircraft to safely perform tasks.

[0004] The technical solution adopted by the present invention to solve its technical problems is: to provide a degradation prediction method for aviation equipment in multiple working conditions and multiple stages, including the following steps:

[0005] Fuse the performance data of aviation equipment under multi-dimensional time-varying working conditions into comprehensive health index data;

[0006] Perform heuristic change-point dynamic detection on the comprehensive health index data, and intercept degradation feature data from the comprehensive health index data according to the detection results;

[0007] Based on a complex non-linear Wiener degradation model, construct a degradation trajectory model of aviation equipment in combination with the degradation feature data, and use the degradation trajectory model to predict the degradation trend of the equipment.

[0008] The step of fusing the performance data of aviation equipment under multi-dimensional time-varying working conditions into comprehensive health index data specifically includes:

[0009] Fuse multi-dimensional time-varying working condition data into a one-dimensional working condition state index;

[0010] Perform standardized preprocessing on the performance characteristic indexes under the working condition state index to obtain standard performance indexes;

[0011] Smoothing the standard performance index into a health index by the exponentially weighted moving average method;

[0012] Calculating the Euclidean distance between the health index and the initial health index using the Euclidean distance to obtain a comprehensive health index.

[0013] The calculation method of the health index is as follows: where HIs t is the health index at time t, SPIs t is the standard performance index at time t, SPIs0 is the standard performance index at the initial time, and α represents the exponential smoothing coefficient.

[0014] Performing heuristic change-point dynamic detection on the comprehensive health index data, and intercepting degradation feature data from the comprehensive health index data according to the detection result, specifically including:

[0015] Intercepting comprehensive health index data of a preset length, and using the kernel density estimation method to fit the intercepted comprehensive health index data to obtain a probability density curve;

[0016] Calculating the extreme points of the probability density curve, and using the last escape rule based on the extreme points to determine the change points of the intercepted comprehensive health index data;

[0017] Continuing to intercept comprehensive health index data of a preset length based on the change points, and repeating the above steps until all comprehensive health index data are exhausted, and taking the finally obtained change points as the elbow points between the stable data and the degradation data;

[0018] Determining the comprehensive health index data after the elbow point as the degradation feature data.

[0019] Based on the complex non-linear Wiener degradation model, constructing a degradation trajectory model of the aviation equipment in combination with the degradation feature data, and predicting the degradation trend of the equipment using the degradation trajectory model, specifically including:

[0020] Constructing a complex non-linear Wiener degradation model;

[0021] Taking the first 70% of the degradation time period in the degradation feature data as the training data;

[0022] Training the complex non-linear Wiener degradation model using the training data; during training, using the log-likelihood function constructed by the maximum likelihood estimation method to estimate the parameters of the complex non-linear Wiener degradation model, and updating the parameters of the complex non-linear Wiener degradation model based on the model parameter estimation values and the drift coefficient. After training, a degradation trajectory model of the aviation equipment is obtained;

[0023] Input the degraded feature data into the degradation trajectory model of the aviation equipment to obtain the prediction result of the equipment's degradation trend.

[0024] The technical solution adopted by the present invention to solve its technical problems is to provide a multi-condition and multi-stage aviation equipment degradation prediction device, including:

[0025] A fusion module, configured to fuse the performance data of the aviation equipment under multi-dimensional time-varying conditions into comprehensive health index data;

[0026] An intercept module, configured to perform heuristic change-point dynamic detection on the comprehensive health index data, and intercept degraded feature data from the comprehensive health index data according to the detection result;

[0027] A prediction module, configured to construct a degradation trajectory model of the aviation equipment based on the complex non-linear Wiener degradation model in combination with the degraded feature data, and use the degradation trajectory model to predict the degradation trend of the equipment.

[0028] The fusion module includes:

[0029] A fusion unit, configured to fuse multi-dimensional time-varying condition data into a one-dimensional condition state index;

[0030] A standardization unit, configured to perform standardization preprocessing on the performance characteristic indexes under the condition state index to obtain standard performance indexes;

[0031] A smoothing unit, configured to smooth the standard performance indexes into health indexes by the exponentially weighted moving average method;

[0032] A distance calculation unit, configured to calculate the Euclidean distance between the health index and the initial health index by using the Euclidean distance to obtain the comprehensive health index.

[0033] The intercept module includes:

[0034] A fitting unit, configured to intercept comprehensive health index data of a preset length, and fit the intercepted comprehensive health index data by using the kernel density estimation method to obtain a probability density curve;

[0035] A calculation and determination unit, configured to calculate the extreme points of the probability density curve, and determine the change points of the intercepted comprehensive health index data by using the last escape rule based on the extreme points;

[0036] A repetition unit, configured to continue to intercept comprehensive health index data of a preset length based on the change points, and repeat the execution steps of the above-mentioned fitting unit and calculation and determination unit until all the comprehensive health index data are exhausted, and use the finally obtained change points as the elbow points between the stable data and the degraded data;

[0037] A determination unit for determining the comprehensive health index data after the elbow point as degradation feature data.

[0038] The prediction module includes:

[0039] A construction unit for constructing a complex non-linear Wiener degradation model;

[0040] A division unit for taking the first 70% of the degradation time period in the degradation feature data as training data;

[0041] A training unit for training the complex non-linear Wiener degradation model with the training data; during training, using the log-likelihood function constructed by the maximum likelihood estimation method to estimate the parameters of the complex non-linear Wiener degradation model, and updating the parameters of the complex non-linear Wiener degradation model based on the model parameter estimation values and the drift coefficient. After training, a degradation trajectory model of the aviation equipment is obtained;

[0042] A prediction unit for inputting the degradation feature data into the degradation trajectory model of the aviation equipment to obtain a prediction result of the degradation trend of the equipment.

[0043] The technical solution adopted by the present invention to solve its technical problems is: to provide a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned multi-condition and multi-stage aviation equipment degradation prediction method are implemented.

[0044] Beneficial effects

[0045] Due to the adoption of the above technical solution, compared with the prior art, the present invention has the following advantages and positive effects: by constructing a health index construction model under time-varying working conditions, fusing the performance data of aviation equipment under multiple working conditions into comprehensive health index data, intercepting degradation feature data from the health index data by constructing a heuristic change-point dynamic monitoring model for multi-stage health index data, and constructing a degradation trajectory model for each unit of the aviation equipment based on the complex non-linear Wiener degradation model, and then predicting the degradation trend of the equipment, thus realizing the accurate prediction of the degradation situation of the aviation equipment. Brief description of the drawings

[0046] Figure 1 is a flowchart of the multi-condition and multi-stage aviation equipment degradation prediction method according to the first embodiment of the present invention;

[0047] Figure 2 is an evolution diagram of the comprehensive health index data in the first embodiment of the present invention;

[0048] Figure 3 is a schematic diagram of degradation data based on change-point detection in the first embodiment of the present invention;

[0049] Figure 4 It is a prediction result graph of the degradation trend of the aviation equipment in the first embodiment of the present invention. Specific embodiments

[0050] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims of this application.

[0051] The first embodiment of the present invention relates to a method for predicting the degradation of aviation equipment in multiple working conditions and multiple stages, as Figure 1 shown, including the following steps:

[0052] Step 1, fuse the performance data of the aviation equipment under multi-dimensional time-varying working conditions into comprehensive health index data. As Figure 2 shown, this step is specifically: fuse the multi-dimensional time-varying working condition data into a one-dimensional working condition state index; perform standardization preprocessing on the performance characteristic indexes under the working condition state index to obtain standard performance indexes; smooth the standard performance indexes into health indexes by the exponentially weighted moving average method; calculate the Euclidean distance between the health index and the initial health index by using the Euclidean distance to obtain the comprehensive health index.

[0053] This step can be implemented by constructing a health index construction model under time-varying working conditions. The input of this model is: time-varying operating conditions Cs and performance indexes PIs, and the output is: comprehensive health index HI. The process is as follows:

[0054] 1. Select J performance index characteristics PIs with obvious trends;

[0055] 2. Encode the multi-dimensional time-varying operating conditions Cs into a one-dimensional working condition state index CI

[0056] 3. for each group of working conditions CI ∈ CI do

[0057] 4. Select the performance indexes PIs under the working condition state index CI

[0058] 5. Fit PIs through a standardization tool

[0059] 6. Transform PIs into standard performance indexes SPIs

[0060]

[0061] where, μ PIs the mean value of the performance index data, σ PIs the standard deviation of the performance index data.

[0062] 7. Smooth the standard performance indicators (SPIs) into health indicators (HIs) by the exponentially weighted moving average method

[0063]

[0064] where SPIs t is the standard performance indicator at time t, SPIs0 is the standard performance indicator at the initial time, and α represents the exponential smoothing coefficient.

[0065] 8. end for

[0066] 9. for each unit i ∈ Z + do

[0067] 10. Calculate the Euclidean distance between the health indicator HIs and the initial health indicator as the comprehensive health indicator HI

[0068]

[0069] where HIs j,t is the j-th health indicator at time t, and HIs j,0 is the j-th initial health indicator at time t.

[0070] 11. endfor

[0071] Return the comprehensive health indicator HI for each unit.

[0072] Step 2: Perform heuristic change-point dynamic detection on the comprehensive health indicator data, and intercept the degradation feature data from the comprehensive health indicator data according to the detection results. As Figure 3 shown, this step is specifically as follows: Intercept the comprehensive health indicator data of a preset length, and use the kernel density estimation method to fit the intercepted comprehensive health indicator data to obtain a probability density curve; calculate the extreme points of the probability density curve, and use the last exit rule based on the extreme points to determine the change points of the intercepted comprehensive health indicator data; continue to intercept the comprehensive health indicator data of a preset length based on the change points, and repeat the above steps until all the comprehensive health indicator data is exhausted, and use the last obtained change point as the elbow point between the stationary data and the degradation data; determine the comprehensive health indicator data after the elbow point as the degradation feature data.

[0073] In this step, the heuristic change-point dynamic detection of the comprehensive health indicator data can be realized by constructing a heuristic change-point dynamic monitoring model. The input of this model is: the comprehensive health indicator data HI and the length L of the health indicator data HI HI , and the output is: the elbow point X between the stationary data and the trend data e . Its process is as follows:

[0074]

[0075]

[0076] Return the elbow point X of all units e 。

[0077] Step 3: Based on the complex non-linear Wiener degradation model, construct a degradation trajectory model of the aviation equipment in combination with the degradation characteristic data, and use the degradation trajectory model to predict the degradation trend of the equipment. This step specifically includes: constructing a complex non-linear Wiener degradation model; using the first 70% of the degradation time period in the degradation characteristic data as training data; training the complex non-linear Wiener degradation model with the training data; during training, using the log-likelihood function constructed by the maximum likelihood estimation method to estimate the parameters of the complex non-linear Wiener degradation model, and updating the parameters of the complex non-linear Wiener degradation model based on the model parameter estimation values and the drift coefficient. After training is completed, obtain the degradation trajectory model of the aviation equipment; input the degradation characteristic data into the degradation trajectory model of the aviation equipment to obtain the prediction result of the degradation trend of the equipment.

[0078] This step can be achieved by constructing a degradation prediction model based on the complex Wiener model, with its input being: time-varying operating conditions Cs and performance indicators PIs, and its output being: degradation trend prediction DT. The process is as follows:

[0079] 1. Obtain the comprehensive health index data HI through the algorithm in Step 1

[0080] 2. Obtain the elbow point X of each unit through Algorithm 2 in Step 2 e

[0081] 3. Intercept the comprehensive health index data HI after the elbow point X, and transform the elbow point X e into the relative zero point to obtain the degradation index data DI e

[0082] 4. Construct a complex non-linear Wiener degradation model

[0083] y i,j =λ i Λ(t i,j ;θ)+σ B B(t i,j )+ε i,j

[0084] where y i,j is the output of the complex non-linear Wiener degradation model, Λ(t i,j ;θ) represents the degradation function, and B(t i,j ​) represents the standard Brownian motion, t i,j represents the monitoring time of the device, θ represents the parameter of the degradation function, λ i represents the drift coefficient, σ B represents the diffusion coefficient, ε i,j represents the measurement error.

[0085] 5. Use the first 70% of the degradation index data DI as the model training data to obtain the degradation increment data ΔY

[0086] 6. Based on the maximum likelihood estimation method, construct the log-likelihood function for the model parameter Θ

[0087]

[0088] where, m i represents the number of times the device is monitored, n represents the number of devices, represents the variance of the degradation data, represents the increment of the device monitoring data, μ λ represents the mean of the drift coefficient, represents the increment of the degradation function.

[0089] 7. Optimize the log-likelihood function based on the simplex algorithm to obtain the estimated value of the model parameter

[0090] 8. for each cell i ∈ Z + do

[0091] 9. Update the drift coefficient of the cell based on the training data of the cell

[0092]

[0093] where, μ λ,p represents the updated value of the mean of the drift coefficient, Δy i,p represents the increment of the device monitoring data, Σ i,p represents the variance of the degradation data, Δν i,p represents the increment of the degradation function, σ λ represents the standard deviation of the drift coefficient, μ λ represents the mean of the drift coefficient, σ λ,p represents the updated value of the standard deviation of the drift coefficient.

[0094] 10. Based on the estimated value of the model parameter and the drift coefficient obtain the degradation trajectory model of the cell

[0095]

[0096] where, Represents the predicted value of the degraded data, Represents the estimated value of the degradation function parameters, Represents the estimated value of the variance of the degraded data, Represents the estimated value of the drift coefficient, Represents the estimated value of the diffusion coefficient, Ω i,p Represents the variance matrix of the Brownian motion, Represents the estimated value of the measurement error coefficient, F i,p Represents the variance matrix of the measurement error.

[0097] 11. Based on the cell-based degradation trajectory model, predict the degradation trend DT of the cell

[0098] 12. end for

[0099] 13. return the degradation trend DT of each cell (see Figure 4 ).

[0100] It is not difficult to find that the present invention constructs a health index construction model under time-varying working conditions, fuses the performance data of aviation equipment under multiple working conditions into comprehensive health index data, constructs a heuristic change-point dynamic monitoring model for multi-stage health index data, intercepts degradation characteristic data from the health index data, constructs a degradation trajectory model for each cell of the aviation equipment based on the complex non-linear Wiener degradation model, and then predicts the degradation trend of the equipment, thereby achieving an accurate prediction of the degradation situation of the aviation equipment.

[0101] The second embodiment of the present invention relates to a multi-condition multi-stage aviation equipment degradation prediction device, including:

[0102] A fusion module, configured to fuse the performance data of the aviation equipment under multi-dimensional time-varying working conditions into comprehensive health index data;

[0103] An interception module, configured to perform heuristic change-point dynamic detection on the comprehensive health index data, and intercept degradation characteristic data from the comprehensive health index data according to the detection result;

[0104] A prediction module, configured to construct a degradation trajectory model of the aviation equipment based on the complex non-linear Wiener degradation model in combination with the degradation characteristic data, and predict the degradation trend of the equipment by using the degradation trajectory model.

[0105] The fusion module includes:

[0106] A fusion unit, configured to fuse multi-dimensional time-varying working condition data into a one-dimensional working condition state index;

[0107] A standardization unit, configured to perform standardization preprocessing on the performance characteristic indexes under the working condition state index to obtain standard performance indexes;

[0108] A smoothing unit, configured to smooth the standard performance index into a health index by the exponentially weighted moving average method;

[0109] A distance calculation unit, configured to calculate the Euclidean distance between the health index and the initial health index by using the Euclidean distance to obtain a comprehensive health index.

[0110] The calculation method of the health index is as follows: where HIs t is the health index at time t, SPIs t is the standard performance index at time t, SPIs0 is the standard performance index at the initial time, and α represents the exponential smoothing coefficient.

[0111] The intercepting module includes:

[0112] A fitting unit, configured to intercept comprehensive health index data of a preset length and fit the intercepted comprehensive health index data by using the kernel density estimation method to obtain a probability density curve;

[0113] A calculation and determination unit, configured to calculate the extreme points of the probability density curve and determine the change points of the intercepted comprehensive health index data by using the last escape rule based on the extreme points;

[0114] A repeating unit, configured to continue intercepting comprehensive health index data of a preset length based on the change points and repeat the execution steps of the above fitting unit and calculation and determination unit until all the comprehensive health index data are exhausted, and use the finally obtained change points as the elbow points between the stable data and the degradation data;

[0115] A determination unit, configured to determine the comprehensive health index data after the elbow points as degradation characteristic data.

[0116] The prediction module includes:

[0117] A construction unit, configured to construct a complex non-linear Wiener degradation model;

[0118] A partitioning unit, configured to use the first 70% of the degradation time period in the degradation characteristic data as training data;

[0119] A training unit, configured to train the complex non-linear Wiener degradation model by using the training data; during training, use the log-likelihood function constructed by the maximum likelihood estimation method to estimate the parameters of the complex non-linear Wiener degradation model, and update the parameters of the complex non-linear Wiener degradation model based on the model parameter estimation values and the drift coefficient. After training is completed, a degradation trajectory model of the aviation equipment is obtained;

[0120] A prediction unit for inputting the degraded feature data into a degradation trajectory model of the aviation equipment to obtain a prediction result of the degradation trend of the equipment.

[0121] The third embodiment of the present invention relates to a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the multi-condition and multi-stage aviation equipment degradation prediction method of the first embodiment are implemented.

[0122] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0123] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the processes and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0124] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction method, and the instruction method implements the specified functions in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0125] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the specified functions in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0126] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims described above.

Claims

1. A degradation prediction method for aviation equipment with multiple working conditions and multiple stages, characterized in that, It includes the following steps: Fuse the performance data of aviation equipment under multi-dimensional time-varying working conditions into comprehensive health index data; Perform heuristic change-point dynamic detection on the comprehensive health index data, and intercept degradation feature data from the comprehensive health index data according to the detection results; Based on the complex non-linear Wiener degradation model, construct a degradation trajectory model of the aviation equipment in combination with the degradation feature data, and use the degradation trajectory model to predict the degradation trend of the equipment.

2. The multi-condition and multi-stage aviation equipment degradation prediction method according to claim 1, wherein The step of fusing the performance data of aviation equipment under multi-dimensional time-varying working conditions into comprehensive health index data specifically includes: Fuse multi-dimensional time-varying working condition data into a one-dimensional working condition state index; Perform standardized preprocessing on the performance characteristic indexes under the working condition state index to obtain standard performance indexes; Smooth the standard performance indexes into health indexes by the exponentially weighted moving average method; Calculate the Euclidean distance between the health index and the initial health index by the Euclidean distance calculation method to obtain the comprehensive health index.

3. The multi-condition and multi-stage degradation prediction method for aviation equipment according to claim 2, wherein, The calculation method of the health index is as follows: Among them, HIs t is the health index at time t, and SPIs t is the standard performance index at time t, SPIs0 is the standard performance index at the initial time, and α represents the exponential smoothing coefficient.

4. The multi-condition and multi-stage aviation equipment degradation prediction method according to claim 1, characterized in that, The step of performing heuristic change-point dynamic detection on the comprehensive health index data and intercepting degradation feature data from the comprehensive health index data according to the detection results specifically includes: Intercept comprehensive health index data of a preset length, and use the kernel density estimation method to fit the intercepted comprehensive health index data to obtain a probability density curve; Calculate the extreme points of the probability density curve, and use the last escape rule based on the extreme points to determine the change points of the intercepted comprehensive health index data; Continue to intercept comprehensive health index data of a preset length based on the change points, and repeat the above steps until all the comprehensive health index data is exhausted, and use the finally obtained change points as the elbow points between the stable data and the degradation data; Determine the comprehensive health index data after the elbow points as the degradation feature data.

5. The multi-condition and multi-stage aviation equipment degradation prediction method according to claim 1, wherein The step of constructing a degradation trajectory model of the aviation equipment in combination with the degradation feature data based on the complex non-linear Wiener degradation model and using the degradation trajectory model to predict the degradation trend of the equipment specifically includes: Construct a complex non-linear Wiener degradation model; Use the first 70% of the degradation time period in the degradation feature data as training data; Use the training data to train the complex non-linear Wiener degradation model; during training, use the log-likelihood function constructed by the maximum likelihood estimation method to perform parameter estimation on the complex non-linear Wiener degradation model, and update the parameters of the complex non-linear Wiener degradation model based on the model parameter estimation values and the drift coefficient. After training, obtain the degradation trajectory model of the aviation equipment; Input the degradation feature data into the degradation trajectory model of the aviation equipment to obtain the prediction result of the degradation trend of the equipment.

6. A multi-condition and multi-stage degradation prediction device for aviation equipment, characterized in that It includes: A fusion module for fusing the performance data of aviation equipment under multi-dimensional time-varying working conditions into comprehensive health index data; An interception module for performing heuristic change-point dynamic detection on the comprehensive health index data and intercepting degradation feature data from the comprehensive health index data according to the detection results; A prediction module, configured to construct a degradation trajectory model of an aviation device based on a complex non-linear Wiener degradation model in combination with the degradation feature data, and predict the degradation trend of the device by using the degradation trajectory model.

7. The multi-condition and multi-stage aviation equipment degradation prediction device according to claim 6, wherein, The fusion module includes: A fusion unit, configured to fuse multi-dimensional time-varying working condition data into a one-dimensional working condition state index; A normalization unit, configured to perform pre-normalization processing on the performance feature indexes under the working condition state index to obtain standard performance indexes; A smoothing unit, configured to smooth the standard performance indexes into health indexes by using the exponentially weighted moving average method; A distance calculation unit, configured to calculate the Euclidean distance between the health indexes and the initial health indexes by using the Euclidean distance to obtain comprehensive health indexes.

8. The multi-condition multi-stage aviation equipment degradation prediction device according to claim 6, wherein, The truncation module includes: A fitting unit, configured to truncate comprehensive health index data of a preset length, and perform fitting on the truncated comprehensive health index data by using the kernel density estimation method to obtain a probability density curve; A calculation and determination unit, configured to calculate extreme points of the probability density curve, and determine change points of the truncated comprehensive health index data based on the last escape rule with the extreme points as a reference; A repetition unit, configured to continue to truncate comprehensive health index data of a preset length based on the change points, and repeat the execution steps of the fitting unit and the calculation and determination unit until all the comprehensive health index data are exhausted, and use the finally obtained change points as the elbow points between the stable data and the degradation data; A determination unit, configured to determine the comprehensive health index data after the elbow points as the degradation feature data.

9. The multi-condition and multi-stage aviation equipment degradation prediction device according to claim 6, wherein, The prediction module includes: A construction unit, configured to construct a complex non-linear Wiener degradation model; A division unit, configured to use the first 70% of the degradation time period in the degradation feature data as training data; A training unit, configured to train the complex non-linear Wiener degradation model by using the training data; during training, perform parameter estimation on the complex non-linear Wiener degradation model by using a log-likelihood function constructed by the maximum likelihood estimation method, and update the parameters of the complex non-linear Wiener degradation model based on the model parameter estimation values and the drift coefficient. After training is completed, a degradation trajectory model of the aviation device is obtained; A prediction unit, configured to input the degradation feature data into the degradation trajectory model of the aviation device to obtain a prediction result of the degradation trend of the device.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the steps of the multi-condition and multi-stage aviation device degradation prediction method according to any one of claims 1-5 are implemented.

Images